| Risk factors and carotid atherosclerosis in hypertensive and control subjects | ||
|---|---|---|
| Prev | Chapter 4. Subjects and methods | Next |
The procedure for blood pressure measurement was in agreement with the recommendations of the American Society of Hypertension (1992). All the blood pressure measurements were recorded with an automatic oscillometric blood pressure recorder (Dinamap, Critikon Ltd., Ascot, UK), which has been extensively validated (Hutton et al. 1984, Silas et al. 1980). The resting blood pressure was measured three times at one-minute intervals on the right arm after the patient had been seated for at least 5 minutes. After the three measurements, the patient stood up, and two blood pressure measurements at one-minute intervals were recorded with the patient standing. The mean value of the three sitting blood pressure measurements was used in the analyses.
Height was measured to the nearest centimeter without shoes. Weight was measured to the nearest 0.1 kg on a lever balance with the subject in light underwear without shoes. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (in meters) squared. Waist circumference was measured to the nearest 0.5 cm with a tape measure midway between the lower rib margin and the iliac crest in light expirium. Hip circumference was measured at the point yielding the maximum circumference over the buttocks. All the measurements were performed by the same specially trained nurses. Alcohol consumption was determined in an interview with a specially trained physician using questions based on the method of Khavari and Farber (1978). Alcohol intake was calculated by summing the average amount of absolute alcohol in the different beverages and expressing the result in grams of absolute alcohol per week. Smoking history was obtained by the questionnaire used in WHO MONICA (Rose et al. 1982) study. The life-time smoking burden was calculated as pack-years (1 pack-year = 20 cigarettes smoked/day in one year). Physical activity was assessed using the method described by Grimby (1986). Women were classified as postmenopausal if at least six months had passed after their latest menstruation and no other relevant reason for the cessation of the menstruation cycle could be identified. A conventional 12-lead electrocardiogram was recorded and analyzed according to the Minnesota code (Prineas et al. 1982).
Carotid artery ultrasonography was performed on every study subject during a separate visit 6–12 months later.
A duplex ultrasound system with a 7.5 MHz scanning frequency in B-mode, pulsed-Doppler mode and colour mode was used according to the same protocol by a single trained radiologist blinded for the presence or absence of hypertension. The ultrasonographic assessment of carotid arteries was performed while the subject was in a supine position with his head turned away from the sonographer at an angle of 45°. Each carotid system was imaged in anterior oblique and lateral planes, transversally and longitudinally. The examiner consistently aimed for the clearest image of the near and far wall of the carotid arteries. The scan head was kept perpendicular to the arterial walls, and the transducer laterally angled to give optimal visualization. The Doppler mode was used to identify the vessels and to evaluate flow disturbances. Each scan of the common carotid artery began just above the clavicle and was moved cephalad past the bifurcation and along both the internal and external branches as far distally as possible. The whole scanning procedure was recorded on a Super-VHS videocassette recorder (Panasonic). All measurements were performed about one year later from the video image on the monitor of the ultrasound device using its electronic calipers.
The structures measured were intima-media thickness (IMT) and the size and number of atheromatous plaques. IMT was defined as the distance between the media-adventitia interface and the lumina-intima interface. IMT was measured at five locations on each side, both on the near and the far wall, i.e. at a total of 20 sites: the internal (ICA) carotid artery about 1 cm distal from the flow divider, the bifurcation enlargement (BIF) and three locations of the common (CCA) carotid artery: proximal, middle and distal at about 1–1.5 cm intervals, depending on the length of the vessel, and the most cranial measuring point was about 1 cm proximal from the bifurcation. The examiner searched for the thickest point of IMT for measurement at each site, avoiding, however, site with atheromatous plaque. The IMT was measured with the instrument´s electronic calipers to the nearest 0.1 mm. IMTs for each patient were calculated as the mean and maximal values.
Arterial plaque was defined as an echogenic structure encroaching into the vessel lumen with a distinct area, resulting in IMT more than 50% greater compared to the neighbouring sites. The numbers and locations of plaques were recorded. The size of each plaque was measured as the maximal diameter in three perpendicular projections: thickness, length and width.
The intra- and interobserver reproducibility of the IMT measurement was assessed in 31 randomly selected subjects (10 men of age > 57 years, 11 men < 43 years and 10 women > 57 years). The repeat measurements were performed from the videotapes 1.5 years after the examination without knowledge of the original results. Variability was estimated by using the mean ± SD absolute difference between paired measurements. The intrareader variability and the correlation coefficient for the mean IMT (CCA/BIF/ICA) were 3% and 0.97 (Pearson’s coefficient) and those for the maximal IMT 9.9% and 0.94. Correspondingly, the interreader variability and correlation were 7.2% and 0.93 (mean mode) and 12.8% and 0.92 (maximal mode).
A wide range of routine laboratory analyses were conducted. After fasting blood had been drawn, the subjects were given a 75-g glucose load. Both 1-hour and 2-hour glucose and insulin concentrations were determined, except for previously known insulin-treated diabetics. The glucose concentrations were measured with the glucose dehydrogenase method (Diagnostica, Merck, Darmstadt, Germany). The serum insulin levels were measured using a two-site immunoenzymometric assay (AIA-PACK IRI, Tosoh Corp., Tokyo, Japan). The C-peptide levels were measured using double-antibody 125I radioimmunoassay (Diagnostic Products Corporation, Los Angeles, CA). The alanine aminotransferase (ALAT) levels and gamma-glutamyl transpeptidase (γ -GT) were measured with the recommended method according to the European Committee for Clinical Laboratory Standards (kits of Boehringer Mannheim GmbH, Mannheim, Germany, catalogue nos. 487368 and 1442546, respectively), using a BM/Hitachi 911 Automatic Analyzer (Naka Works, Hitachi Ltd., Ibaraki-Ken, Japan) and a Monarch 2000 Chemistry System, Instrumentation Laboratory Inc., Lexington, USA, respectively. The uric acid levels were measured with the enzymatic method (Kone Instruments no. 981030, Espoo, Finland) using a Monarch 2000 Chemistry System, Instrumentation Laboratory Inc., Lexington, USA. Very low density lipoprotein (VLDL, d < 1.006 g/ml) was isolated by spinning plasma in a Kontron TFT 45.6 rotor at 105,000 g and 15ºC for 18 h. A half milliliter of the VLDL-free fraction was mixed with 25 µl of 2.8% (w/v) heparin and 25 µl of 2 M manganese chloride. After centrifugation at 1000 g and 4ºC for 30 min, aliquots of the supernatant were taken for an analysis of the high-density lipoprotein (HDL) concentration. The low-density lipoprotein (LDL) concentration was then calculated by subtracting the cholesterol concentration in HDL from that in the VLDL-free fraction. The concentrations of total cholesterol and triglycerides in the plasma and lipoprotein fractions were determined by enzymatic colorimetric methods (kits of Boehringer Diagnostica, Mannheim GmbH, Germany, catalogue nos. 236691 and 701912), respectively, using a Kone Specific, Selective Chemistry Analyser (Kone Instruments, Espoo, Finland).
The data were analysed with the Statistical Analysis System, version 6.08 (SAS Inc., Cary, North Carolina, USA) on a VAX computer. The data are presented as means with SD, or SE of the mean unless otherwise stated. The Chi-square test was used to test the differences in frequencies. Spearman’s and Pearson’s correlation coefficients were used when estimating the correlation between variables. Analysis of variance was used in the comparison of more than two groups, and the adjustment for confounding factors was performed using the analysis of covariance in the general linear model or the regression model (Armitage & Berry 1987). Appropriate precautions in executing and interpreting the multivariate methods, including tests for interactions and for the stability of coefficients, were observed in accordance with a review (Concato et al. 1993). When the number of subjects in one or more classes was small, the non-parametric equivalent for ANOVA (Kruskal-Wallis’ test) was used. If a significant difference in ANOVA was obtained, the Bonferroni correction was applied. Since the lipid values and many other variables are associated with each other, no further corrections for multiple comparisons were made, to avoid overcorrection (Rothman 1990). Unpaired t-test for independent samples was used in the comparisons of two groups. Multiple logistic regression analyses were performed to investigate the associations of the different variables with the metabolic syndrome. The CIA program (Gardner & Altman 1989) was used to calculate the confidence intervals when appropriate. In all the statistical analyses, logarithmic values for particularly skewed variables were used, including total and VLDL triglycerides, GTP and all insulin values. p-values < 0.05 were considered statistically significant.
To control the effect of confounding factors on dependent variables, adjustments of two types were utilized. To adjust the mean value of a particular group, analysis of covariance was used. Further, adjustments based on linear regression (Siervogel et al. 1980) were applied to get an adjusted value for a given variable for every subject. In that case, the following formula was used:
Pa = P - ab (b – B) - ac (c – C) - ad (d – D) - ...;
where Pa = individual’s adjusted value of a parameter; P = Individual’s observed value of a parameter; ab, ac, ad, ... = sex-specific regression coefficient of a parameter on covariate b, c, d,...; b, c, d, ... = individual’s value of covariate and B, C, D, ... = sex-specific mean value of covariate. The whole male or female (control) population was used as a source of regression coefficients. Due to the marked gender differences in many variables, such as lipid and lipoprotein values, alcohol consumption and smoking status, the adjustments and statistical analyses throughout the present study were performed separately for men and women, except in pooled analyses, gender being a covariate in the analysis of covariance.